| Forum: Knowledge Base |
|---|
| Topic: HyperVM Manual |
|---|
|
| Topic: CentOS OpenSSL update crashes lxlighttpd (Bug #68) |
|---|
|
| Topic: HyperVM shows 2ips, but Lxadmin shows only one. |
|---|
| HyperVM shows 2ips, but Lxadmin shows only one. [message #6260] |
Fri, 08 June 2007 12:53 |
Lxhelp
Messages: 23691
Registered: July 2006 |
The Champion |
|
|
We have had a peculiar problem, where customer would see that the hypervm shows that there are two ips configured on the system, but lxadmin will only notice one or none.
The problem is actually with hyperVM, but it is not a bug. What happens is this. When hyperVM adds an ip, it will check whether it has already been assigned by pinging it, and will not allow you to add an ip if the ip is already assigned to some other hosts. Now in this case what has happened is that, when you added the ipaddress, the ip was free, but after the addition, the ip was allocated to some other host. Now when you reboot the vps, openvz will refuse to use the ip because it is being used by someone else.
Please note that this is not a bug.
If you are seeing this problem, just ping the ip that's not appearing in Lxadmin, and you will see that it will respond.
The solution is to contact your DC and ask them to reassign this ip address.
Thanks.
|
|
|
| Topic: Find the VPSID from the process ID |
|---|
| Find the VPSID from the process ID [message #4861] |
Mon, 02 April 2007 13:09 |
|
If you want to find which VPSID is running a particular process, find the process ID from "top" and use the following command to find it out:
vzpid <pid>
example:
[root@venus ~]# vzpid 29434
Pid VPSID Name
29434 3330 smaug
|
|
|
| Topic: Number of ipaddresses quota variable |
|---|
| Number of ipaddresses quota variable [message #3819] |
Mon, 19 February 2007 14:05 |
Lxhelp
Messages: 23691
Registered: July 2006 |
The Champion |
|
|
We have finally solved all the issues with ipaddress, by properly introducing the number of ipaddresses' variable. You can set this inside both the vps and the client, and like everything else, it is hierarchical. If you have clients already added to the system, you will have to go to 'client home' -> 'limit' and set the 'number of ipaddresses' to a proper value. You will have to set it for Plans too, and if you don't, then a default of '2' will be used, when a vps/client is created using that particular plan.
If you are used to the HyperVM herarchical quota system, all you need to know is that we have introduced another quota variable called as the 'number of ipaddresses'.
Thanks.
|
|
|
| Topic: Openvz Kernel Updates |
|---|
| Openvz Kernel Updates [message #3785] |
Sun, 18 February 2007 16:06 |
Lxhelp
Messages: 23691
Registered: July 2006 |
The Champion |
|
|
Please update your openvz kernel to the latest version.
up2date --nosig ovzkernel
up2date --nosig ovzkernel-smp
Open grub.conf and switch to it.
Thanks.
|
|
|
| Topic: Centos domU on Fc6 dom0, cpu utilizatio too high. |
|---|
| Centos domU on Fc6 dom0, cpu utilizatio too high. [message #3770] |
Sat, 17 February 2007 16:34 |
Lxhelp
Messages: 23691
Registered: July 2006 |
The Champion |
|
|
There's a minor problem with centOS domU on Fc6 dom0 leading to very high cpu utilization by domU owing to the redhat nash shell, which is spawned by the initrd of generated in fc6 going berserk. The problem is that the centOS initscripts doesn't have the logic to kill the nash which is only mean to run inside the initrd. To fix this, just add this line to the domU /etc/rc.local file, once it comes up.
/etc/rc.local
This will ensure that nash will be forcibly killed. Without this, this one DomU can completely hog the CPU and very badly affect the performance.
Thanks.
|
|
|
| Topic: Re: Updating from very old versions. The CP disappears. |
|---|
|
| Topic: Update hangs and you get a message "System is updating itself..." |
|---|
| Update hangs and you get a message "System is updating itself..." [message #3267] |
Sun, 28 January 2007 03:09 |
Lxhelp
Messages: 23691
Registered: July 2006 |
The Champion |
|
|
Sometimes it is possible that an update of hyperVM can hang due to some network issue, and this would mean that you won't be able to perform any write actions on the system, since hyperVM will lock itself while updating so as not do anything with the system in that inconsistent state of the code. If you are getting this message, just login to the server and run the commands below:
ps ax | grep update.php
[ get the pid of the update.php process]
kill <pid>
ps ax | grep updatecleanup
[ get the pid of the updatecleanup process]
kill <pid>
However, if you are facing this routinely, then there is something wrong with the system, and please contact our support personnel to look into it.
Thanks.
|
|
|
| Topic: Fine grained client control. |
|---|
| Fine grained client control. [message #3199] |
Wed, 24 January 2007 11:13 |
Lxhelp
Messages: 23691
Registered: July 2006 |
The Champion |
|
|
HyperVM 1.4 now features much more fine grained client access control delivering the hosting provider the highest flexibility in limiting or allowing the privileges of their customers. The most important ones are 'can add vps' and 'can change limit of vps' flags, the functions of which will be explained below.
"Can add vps flag": Once this is unset, the customer will not see either the vps template nor the 'add' buttons in the vps listing. He won't be able to delete the vpses either.
"Can change limit flag": Unsetting this flag prevents the customer from changing the resource limits of the vpses. Merely preventing a customer from adding a vps will not prevent him from redistributing his total quota across to the vpses. But unsetting this will mean that the customer will be limited to just managing the basic functionalities of the vpses and not change the resource usages.
For admin to access all the vps in the system in a single listing, you can go to 'server home' -> 'virtual machines', and you will be able to see all the virtual machines running on that particular server. This means that even though the vpses are owned by the clients, admin can still have a bird's eye view of the every single vps running on the system.
If you have any questions or clarifications, please contact lxhelp . at . lxlabs.com
Thanks.
[Updated on: Wed, 24 January 2007 14:05] by Moderator Report message to a moderator |
|
|
| Topic: Re: Installation using up2date hangs |
|---|
| Re: Installation using up2date hangs [message #3075] |
Wed, 17 January 2007 22:27 |
Lxhelp
Messages: 23691
Registered: July 2006 |
The Champion |
|
|
If your up2date is hanging, then the issue is primarily with the DNS. Check your /etc/resolv.conf and see if the DNSes listed there are valid. Up2date hangs are primarily caused by DNS problem, since up2date's resolver logic seems to be brain damaged. You can easily solve the problem by temporarily installing bind on the server, and adding 127.0.0.1 to your /etc/resolv.conf, and then deleting it once the installation has been successfully completed.
yum -y install bind bind-chroot
service named restart
Add this to /etc/resolv.conf
Run the installer again.
After the installation is complete, remove both bind and the entry from /etc/resolv.conf
Thanks.
|
|
|
HyperVM & Kloxo Support
Members
Search
Help
Register
Login
Home
